We have compiled a large sample of isolated central galaxies from the Sloan Digital Sky Survey, which do not have a neighbour of comparable brightness within a projected distance of 1 Mpc. We use the colours, luminosities and surface brightnesses of satellite galaxies in the vicinity of these objects to estimate their atomic gas content and to derive the average total mass of HI gas contained in satellites as a function of projected radius from the primary. Recent calibrations of merging timescales from N-body simulations are used to estimate the rate at which this gas will accrete onto the primaries. Our estimated accretion rates fall short of those needed to maintain the observed level of star formation in these systems by nearly two orders of magnitude. Nevertheless, there are strong correlations between the total mass of gas in satellites and the colours and specific star formation rates of central galaxies of all stellar masses. The correlations are much weaker if we consider the total stellar mass in the satellites, rather than their total gas mass. We ask why star formation in the central galaxies should be correlated with gas contained in satellites at projected separations of a Mpc or more, well outside the virial radius of the dark matter halos of these systems. We suggest that gas-rich satellites trace an underlying reservoir of ionized gas that is accreted continuously, and that provides the fuel for ongoing star formation in galaxies in the local Universe.